# Dartmouth's 1955 AI proposal began with a wager about what could be described > The document remembered for a name also set down a research gamble. Published 2026-11-08 Canonical: https://machines-on-machines.anth.us/posts/dartmouth-ai-proposal-1955/ ::image{src="assets/dartmouth-ai-proposal-1955/dartmouth-ai-proposal-1955.jpg" alt="The white façade of Dartmouth Hall at Dartmouth College" credit="Image: Kenneth C. Zirkel, CC BY 4.0, via Wikimedia Commons."} Dartmouth Hall in 2024. The photo shows the campus, not the 1956 meeting. [Image source](https://commons.wikimedia.org/wiki/File:Dartmouth_Hall_at_Dartmouth_College.jpg). [Licence](https://creativecommons.org/licenses/by/4.0/). In August 1955, John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon proposed bringing researchers to Dartmouth for two months the following summer. They put a broad conjecture at the center of the plan. If aspects of intelligence could be described precisely enough, a machine could in principle simulate them. The proposal listed problems in language, abstraction, neural networks and self-improvement. Its authors also acknowledged that available machines could be too slow or have too little memory for some tasks. They argued that writing programs that could make good use of the machines was itself a central obstacle. ## An agenda came before the results The surviving document records what the organizers hoped a small group could accomplish. Their conjecture gave researchers a common problem to work on while leaving room for very different approaches. It was a plan for work, with a budget and proposed participants, as well as a claim about what might be possible. Reading the proposal on its own terms preserves that uncertainty. The phrase artificial intelligence gave the project a name. The pages that follow show how much the people proposing it still had to figure out. [Read the source](https://www-formal.stanford.edu/jmc/history/dartmouth/dartmouth.html).